Writing a Job Description with AI: Recruiter Prompts

Writing a job description often takes longer than you’d like: finding the right tone, listing duties without forgetting anything, avoiding phrasing that excludes certain candidates. AI doesn’t replace your knowledge of the role, but it speeds up structuring and improves the clarity of your final text. Here’s how to use it effectively, with tested prompts and a before/after example.

Why use AI for a job description

A poorly written job description costs money: off-target applications, interviews that fall short, unpleasant surprises at hiring. AI helps in three specific areas:

• Structure: it automatically organizes company background, duties, desired profile, and conditions—no section gets missed.
• Clarity: it rephrases vague language (“great interpersonal skills”, “team player”) into more concrete criteria.
• Inclusivity: it spots phrasing that discourages certain candidates without you realizing it.

This work gets done in minutes instead of an hour writing from scratch. The recruiter keeps control of the content; AI handles the form.

The 5 prompts to use

1. Structure a complete first version

“Write a job description for a [job title] in [industry]. Structure it in 5 sections: company introduction, main duties (5-7 points), required skills, desired profile, conditions (contract type, location, salary if known). Professional but accessible tone, 350 words maximum.”

This prompt gives you a solid foundation that you then adjust with details specific to your company.

2. Rephrase in inclusive language

“Here’s a job description: [paste text]. Identify terms or phrasing that might discourage applications from women, people with disabilities, or non-traditional profiles. Suggest a rephrased version without changing the core meaning of the duties.”

AI spots things like implicitly gendered adjectives (“young dynamic team”) or requirements disproportionate to the actual role.

3. Adapt tone to company culture

“Rewrite this job description [paste text] with a more [warm / formal / direct] tone, keeping factual information intact. The company is a [startup / SME / large corporation] in the [X] sector.”

Useful when the first version sounds too generic or doesn’t match your employer brand.

4. Generate multiple job titles

“Suggest 5 different titles for this role: [brief description of duties]. They should be clear for job search engines, free of internal jargon, and appeal to both junior and experienced candidates.”

The job title drives much of your posting’s visibility. Testing several formulations before publishing helps you reach better-matched candidates.

5. Check for bias and length before publishing

“Review this job description [paste text] and flag: overly long sentences, repetitions, requirements that aren’t justified by the role (degree, years of experience), and missing elements for a candidate discovering the posting.”

This final prompt serves as a last-pass review before posting on job boards.

Before/after example

Before (manual writing, 15 minutes, still incomplete):

“We’re looking for a motivated and dynamic client manager to join our team. Good presentation required, team spirit essential. Position to fill urgently.”

This text says nothing about actual duties, salary, or expected skills. It also implicitly discourages certain profiles with “good presentation”.

After (via prompt #1 then #2, 4 minutes):

“Client Manager (B2B) — Join a 6-person team to manage a portfolio of 80 professional accounts. Duties: account follow-up, complaint handling, monthly reporting, participation in renewal negotiations. Profile: comfortable with CRM tools, experience in client relations appreciated (entry-level accepted with on-site training). Permanent contract, Lyon, €28-32k depending on experience.”

The gain isn’t just cosmetic: the second version already filters for better-matched candidates through specific criteria, and it’s ready to publish without further editing.

Making the description truly inclusive, not just “politically correct”

Inclusivity isn’t just writing “M/F” after the job title. It’s verifying that every requirement is genuinely necessary for the role. A useful prompt for this: ask AI to justify each criterion (“why is this degree essential?”). If it can’t find solid justification, that’s often a sign the requirement should be loosened.

This same vigilance applies later in the recruitment process—for example during an annual review with AI or when preparing questions to expect in a job interview from both candidate and recruiter perspectives.

Compare phrasings before publishing

Each AI has its own style: some are more concise, others more formal, still others more creative with job titles. Generating the same description with two or three models lets you spot the version best suited to your employer brand before sharing. It’s also a good way to verify that no bias crept into just one version—as explained in our article on why the same prompt gives different results across different AIs.

For this, you don’t need to juggle tabs and subscriptions: with try noov.ai free, you can submit the same prompt to ChatGPT, Claude, Gemini, or DeepSeek in a single interface and compare the generated descriptions side-by-side before choosing the best version. If you recruit regularly, this comparison habit is worth applying to other HR tasks too, as detailed in our guide on using multiple AIs simultaneously.

In summary

A successful job description with AI always follows the same pattern: a structuring prompt, an inclusive-language review pass, tone adjustment, and final bias checking. The time savings are real, but quality depends on what you feed into the prompt—the more context you provide about the role, the more ready-to-use the output will be.